OSS Capital's Vision for Open Source Software
Now, we've got a category in COSS that's valued at about $220 billion. More than 50 companies have reached over $100 million in revenue each and we've had about eight IPOs over that timeframe.
Check out our open-source Feature Flagging system - Flagsmith on Github! I'd appreciate your feedback ❤️
This week we decided to publish an interview that we did a while back with Joseph "JJ" Jacks who is the Founder and General Partner of OSS Capital. If you aren't familiar with OSS Capital, they are the world's first early-stage COSS (Commercial Open-Source Software) investment fund. They define COSS companies as "Any company that would not exist without the parallel co-existence of a given Open-Source technology".
It was a wide-ranging conversation that touched on some of the more macro trends for the OSS community, the companies that have risen from the approach and where JJ sees things progressing.
Hope you enjoy!
This episode is slightly unusual and it is not speaking directly to a founder engineer. We have Joseph Jacks. His LinkedIn page says COSS missionary which is one of the reasons that we’ve got him on. Do you want to tell us a little bit about you and your mission is?
Sure. Just like SaaS materialized years ago to describe companies like Salesforce, I got super lucky early in my career years ago to work at a different kind of company. One that a lot of software companies self-identify as open core or commercial open source but we’ve never seen language emerge with consistency that describes these companies. Just like SaaS is like, “We have SaaS companies.” There are the SaaS indices, language, metrics, and so on. Companies that would not exist without the parallel coexistence of a given open source available, potentially, technology should be defined as their own category. In the same way that SaaS refers to a particular type of company. That’s why I use this COSS acronym. For me, it only refers to a type of company. It has nothing to do with replacing open source or a different open source or anything.
COSS, Commercial Open Source, refers to any company that would not exist without the parallel coexistence of a given open source, open core technology. This is business model agnostic as a term. You can be selling ice cream, IP, or a cloud service. The characteristics, behaviorally and structurally, of these companies are fundamentally different on many dimensions. I developed a lot of passion for this observation several years ago. That motivated me to start a fund based on this thesis exclusively in late 2018. It’s called OSS Capital. We’re a pre-Series A fund and we also started a conference aiming to educate the world about commercial open source dynamics and also bring together founders that have built successful COSS companies. It’s like MongoDB, Red Hat, Elastic, Cloudera, and HashiCorp. The long list of companies you’ve heard of in that space. This category has materialized through organic means since 2010. In 2010, we only had Red Hat as an example. The whole category was $10 billion in value and Red Hat was 90% of that.
Now, we’ve got a category in COSS that’s valued at about $220 billion. More than 50 companies have reached over $100 million in revenue each and we’ve had about eight IPOs over that timeframe. Something in $5 billion in venture capital poured into this category across all stages over that timeframe as well. 2020 was a record year, about $3.5 billion seed to series F, late-stage. It’s just in 2020 for COSS companies. I can say I’ve been enormously fortunate and lucky to have timed this well on two dimensions. One is this incredible hyper-growth of almost every asset class, but in particular private equity and, to a large extent, more public equity is driving that incredible growth in the technology sector.
In particular, commercial open source has been doing really well. Coronavirus accelerated the rate of growth, adoption, and understanding of the fundamental benefits of remote, distributed in nature being one of the common things. I’m not universally sure in all cases but it propelled this category forward by five years. We’re living 2025-2026, not 2021. Thanks to this double-edged sword known as Coronavirus. I’m a missionary because what we’re doing is, all those data points aside, pretty contrarian and unique. I remain shocked that we’re the only group and firm focused exclusively on this and doing nothing else. That’s why I like to say that we’re missionaries on a multi-decade journey. It goes without saying that every major venture firm and investor out there is extremely keen and interested in this category.
The final thing I’ll say is I’m a missionary because I had zero intent in my career to become an investor or a VC. I don’t like to categorize myself as that at all. I have been deeply involved in the Kubernetes ecosystem since the beginning. I started the first company there. I started the conference around Kubernetes and I got extremely lucky through all that as well. The big distinction is companies that are forming around open source technology, in most cases of open source software technology, are really different. The key thing is that the company would not exist without the coexistence of a core open source project, typically singular.
In some cases, it’s plural like HashiCorp has 3 or 4. It’s a meaningful core technology there. The way the company feels, breathes, and operates has a lot of critical distinctions and differences in many areas. Things like product development, hiring, engineering management, strategy, and business models are very different. We have been learning in public about a lot of this. The conference is very positive, so many inclusive. We invite everyone from the industry to participate in that. We’re category missionaries. We’re doing a lot of other things. If you follow us on Twitter, we announced we’re launching an ETF, Exchange Traded Fund, which will allow retail public investors to gain access to this asset class.
We have six public COSS companies, MongoDB, Elastic, Fastly, Cloudera, JFrog, and Talend, a company I worked at before. This represents a public batch of securities that you can now trade. Interestingly, those six public COSS companies have outperformed the most widely known cloud index. The Bessemer Venture Partners’ EMCLOUD Index over the previous years was equally weighted across those six companies. It’s an exciting category. Most of the growth is still in private equity, so it’s seed Series A, Series B stage companies.
There are less than 2,000 COSS companies on Earth as compared to hundreds of thousands of SaaS companies. A very niche and very nuanced category require a deep understanding of core open-source fundamentals. There’s a decent amount of evolution there in the licensing domain and you can dig into some of that. Overall, I’d say that in my life and my career, having followed intuitions and curiosities, I got extremely lucky.
What do you say to people at parties when they ask you what you do?
I’d give the same answer that I gave you and describe a bunch of different things. The big thing that I would describe as a level set for people that understand tech and understand there are tech companies and software companies. I use this open core, closed core abstract analogy. In my mind, there are two types of software companies in the tech world. On one hand, open core and on the other hand, closed core.
On the open core side, companies that have an open-source or open technology as the basis for the whole company are different in a lot of key ways. Many different distinctions. Closed core companies are also very different in a lot of key ways. I use that to describe why open core companies are interesting on every level across every possible category and vertical. An open core Facebook, Zoom, or Google would look very different than the current reality, which is a closed core, proprietary, black box. The privacy issue is becoming a pretty big deal.
Elon Musk tweeting about Signal is causing their servers to combust spontaneously. No one knows what happened when they had to respond to those 50 million or change new users that joined Signal over a week. I think the world is slowly tipping towards this and understanding more of it but up until now, open-source has only been understood by geeks and software programmers. At parties, I do have to try to simplify, level set, and explain what is open and what is closed. Most people don’t even care. They look at their iPhone, click one of these apps, and use it without discrimination. They don’t know or care if it’s open core or closed core. They just use it. I think that will change over time a lot in society as people understand the implications of their control over the technology versus controlling the technology over them.
I’ve also brought on Matt Althauser, who’s part of the Flagsmith family, for this episode. Matt’s history and background are in closed companies. He joined the team and me at Flagsmith as an investor through Polychrome and also to help with our go-to-market. I thought it would be interesting to hear his input because he’s had a fairly rapid indoctrination into an open core business where his experience is very much the opposite. Matt, what are your feelings at this point in time?
So far, it’s been a really interesting learning curve even at the close core companies where I’ve been over the years. You’ve always heard about and seen open source and what’s going on. Honestly, at the close core companies where I’ve been is we’ve tried to use open sourcing little things almost as a growth or a marketing tactic. It’s interesting to now be in a company where the core IP is open and it isn’t a tactic. It’s a part of our identity as a team. It’s interesting when you look at making a decision at a closed core company and the process you go through, and the same decision that you want to make when you’re going through it as an open core company. It’s very different.
You have to be much more thoughtful from an outside in perspective, as opposed to inside out. We want to deliver a feature and that’s exciting, or we want to innovate and that’s great, but we have to keep the community in mind. That is something that is new but it’s a very exciting part because one of the coolest things that I’ve experienced since being here is when we get for an amazing addition to our product, and it’s just like, “Wow.” You would never get that in a closed core business. That’s been cool to see. One of the things that we, I don’t know if that’s true for you, but grappling with how do you measure the metrics and what are the metrics? What’s the balance between how much should we be pushing people towards our GitHub repo and how much should we be focusing on monetization?
Keeping the growth of those two things in mind at all times are also differences that we’re learning to balance and get right. It’s been really fun. You have to bring more perspectives day-to-day with the decisions you make. I think your thesis is very interesting because if I was talking to an engineer who wanted to start a company, I would literally tell them to go and look at which great SaaS companies don’t have an open core competitor and consider those as very strong alternatives just because of the trust that we can build with our customers right out of the bat. That’s a huge differentiator. It’s been really cool so far. It’s been really fun.
Joe, you touched on an interesting point about people’s understanding of what it means. From your experience, how often you’ve been in a classroom with people from JPMorgan or whatever. Do those guys get it? Do they think of the penguin and people with bad beards?
That’s an interesting question. Classrooms with people from JPMorgan, I’m not an investment banker. The only people I know at JPMorgan are enterprise architects and CTO CIO types. I do spend a lot of time with institutional investors like the LPs venture because we’re a fund and we have to raise money from meaningful people more than just individuals. People that aren’t in the deep technologies, startup world, distributed systems, or hipster programmers, they definitely don’t understand a lot of the nuances, so I definitely spent a lot of time educating. I think more people are understanding it because they see GitLab, Mongo, and MuleSoft went public. GitHub and Red Hat got acquired. Elastic and Cloudera went public. They see all these companies going public and there are commonalities.
On the public market side, when a company goes public in a large quantity that has similarities, this is very in line with asset class formulation. Asset classes are equities that look the same and have similar characteristics. I wouldn’t say COSS is an asset class. It’s more of a category. I was thinking about this like, “Is COSS an asset class?” It’s not really an asset class. It’s a type of software company at the end of the day. It’s a type of tech company effectively. It could be a hardware company or a software company. More people are understanding it. That’s definitely a new phenomenon like at the later stages of money and institutional investing in private equity and public equity or whatever. People en masse are starting to see this because of all the companies going public.
A lot of the late-stage financings and, frankly, people like Mike Volpi at Index paving the way by investing a lot in this category. People like Peter Fenton at Benchmark who has a big thesis here. Neither of those fine gentlemen, I both know reasonably well, focus on this exclusively. They have paved a huge and critical path for lots of different stakeholders in the startup world, business world, and the technology world to understand that companies built around open-source technology can be successful. They don’t all use the same playbook. There are certainly different approaches to different things but the world is definitely starting to wake up to a lot of that momentum, for sure.
I got a chance to work with Peter Fenton when I was at Optimizely. It’s interesting you mentioned him because he’s a big thinker. He’s seeing things way out in front of the curve. I’m not surprised to hear he’s been doing work in the space. I’d be curious to understand as you’ve evolved over the years as a central operator in the Kubernetes space. Now, on the fund side of things, there’s not a playbook, but are you starting to see, at least, early-stage lessons that you can share with teams so that they can execute? I’m sure that’s a big part of your fund, but I’d love to hear any gems or wisdom that you’ve learned along the way from seeing these teams operate early on.
I’ve resisted the temptation because it’s a very significant temptation to go like, “Here’s the playbook. Here’s the framework. Steps A, B, through to Z.” Each COSS company that we’ve studied out of hundreds has taken a different path to whatever your subjective definition of success is. If it’s a revenue milestone, that’s one subjective definition of success. If it’s headcount or fundraising, what have you, but for me, I tend to think more about principles instead of tactics.
Principles are if you make some key commitments, and you set yourself up in a way to operate alongside those principles tightly, then you’ll probably set yourself up for success versus if you buy into some procedural cookbook or playbook, you’ll more likely end up with a survivor bias and prone to getting misled potentially by generalizations. Like, “It worked for these twelve companies and now, we’re going to try the exact same thing.” Each company still needs to think critically about what can work best for them and not necessarily like, “We’re just going to copy and paste these different things.” For principles, I think it is critical for early-stage COSS companies to self-identify. That’s one of the big steps.
I’m not stack ranking anything here. I’m giving you an answer that is off the top of my head. Self-identification and self-awareness are really critical. The biggest principle mistake I see in a lot of early-stage COSS companies is that they copy and paste things that work for SaaS companies at the early stages. SaaS companies, 95% of the time, are closed core. They’re fully proprietary. By the way, there’s absolutely nothing wrong with that. I am not an ideological person. I’m not like, “One is better than the other as far as good or bad.” They’re different approaches and they have different tradeoffs. That’s one thing, self-identification. More companies are self-identifying, so that’s a good sign. I’m starting to see that a lot more, so it’s encouraging.
The other thing that is important is something that you mentioned, Matt and Ben as well, it’s commitment to being positive-sum oriented and building a community that is positive-sum. There’re certainly arguments against this, but I tend to think that closed core companies that build communities are more zero-sum in nature. Think of Twilio or Atlassian are some examples. Again, very successful and large companies. Nothing wrong with them. There’s not any unethical or immoral approach to that dynamic but in terms of their capitalism approach, it tends to be more zero-sum oriented. It’s a stretch to say that, so being a positive-sum in your mindset is really critical and understanding that. That’s another big principle.
Coming from closed core, sometimes you would see businesses that would talk about how much business was built in the ecosystem of their technology, which is this idea of like, “You could build on top of us.” Have you seen people effectively measure and celebrate this positive-sum from what they’re doing as an open core company?
I have seen people celebrate that. I’ve seen companies emphasize that as something that sets them apart or something that enforces and influences their culture, their way of hiring, and self-selecting for team members or strategies, what have you. There’s also the argument independent of this category that positive-sum is the only way to go in the world. Tesla is very positive-sum. Letting any other car manufacturer use their charging network. That’s very positive-sum in nature. Elon is positive-sum but as far as I can tell, every one of his companies is a closed core company, effectively.
Again, nothing wrong with that and I think everyone can be very positive-sum. The thing that excites me a lot about COSS companies and companies that have this open core model is that you’re inherently and fundamentally positive-sum. It’s hard to get away from that. It’s hard to escape those fundamentals. If you try to shift away from them, there are lots of subjective definitions of what that would mean and what that would entail. If you try to resist those fundamentals. You’re going to set yourself up for failure or suboptimal progress. Being positive-sum is important. It’s not just related to go-to-market or competition, raising capital, or anything, it’s holistic. It affects every decision you make almost.
I can understand for certain companies like DoorDash. There’s not much benefit for them in any real dimension to open source their platform, but for companies like Flagsmith, PostHog, developer tools, databases, and infrastructure technology. In the following years, do you think there will be any closed providers of those things?
I’m not a maximalist here as much as I have this name for being a missionary. I believe closed core companies are going to be around for the next 100 years. What we’re starting to see is this COSS category emerging in a different way. The future is a story of and not or. We’re going to see a lot of heterogeneity and both approaches working and succeeding. People will have to make the decisions of like, “What do you want to be a part of? Do you want to be a part of something that’s fundamentally across every dimension, concentrating value capture to one provider, or something that creates abundance?”
We have proof that companies that focus on creating abundance and delivering something of significant value and discover a business model where they capture a small slice of the value, they create can become extremely valuable. They can be worth tens of billions. We’re going to see several 100 plus billion-dollar COSS companies. By the way, I do think it could make sense to have an open core DoorDash or Instacart for different reasons. It’ll be relevant to every part of the stack, not just infrastructure, software, dev tools, distributed systems, or cloud but in applications, consumer software, enterprise apps, SaaS, and everything.
How does that manifest itself? It feels like a huge leap. To persuade a feature flagging or to have a feature flagging company go, “We want to open source it but it makes so much sense to as well.” If you’re running Lyft or Uber, how’d you argue that in a boardroom?
That’s an interesting question. Right now, it’d be hard to argue for that because the characteristics of those businesses are like the world of atoms-oriented. Companies that are very deep in the world of bits and that’s pretty much all they deliver. This open core and this whole COSS model are much more relevant to them. Companies that are deeply in the world of atoms, meaning you have people jumping on a go-kart, transporting a pizza from point A to point B in a metro area, or in a suburban area, or whatever. That’s the fundamental mechanism for measuring how you create value. By the way, Uber Eats is the largest and fastest-growing business in the world over the last years from $0 to $40 billion in revenue.
There are definitely parts of the world that will take a long time to even explore and answer those boardroom questions. Maybe we should question our fundamental design. Right now, I don’t think it’s even remotely relevant to bring those questions up. To answer your question about how this will manifest, I’ve got this question many times, and I could be very wrong about this. For me, the way I extrapolate out to what the tipping point is for all of this open core stuff making sense in many different parts of the stack and the company sphere than infrastructure software is the following. We’ve got about 4 billion humans on the internet. Years ago, we had about 100 million. We’ve seen a lot of growth there. It’s such a finite blink of an eye period of time relative to the existence of our species.
We’ve got about 30 million programmers, people who can talk to the computer in its own language with code and creating all of the tools that have served to enable the whole tech industry. Open source creates all the value for the entire tech industry, software, and even hardware, arguably. People who capture that value our proprietary closed core companies, and we’re starting to see the open core COSS companies to capture some of that value directly. At the low levels of the stack first, but going higher and higher. The materialization trigger or lever for accelerating that rate of open source eating everything is the sheer raw number of programmers and developers.
If you look at the growth rate and you projected out ten years, we’ll probably have 1 billion coders, programmers, and software creators, however you define them. That role and the profession of software engineers is going to evolve a lot. People might not be writing assembly and C and C++ code in 20 to 30 years, but they’ll still be programming in some object-oriented-ish procedural/high-level language, but they’re going to be writing code. When we have about a billion people creating software, and all of that largely being open-source or maybe source available in nature, I personally still think it’ll be very open source. The lever point for every possible business that you could have a hard time imagining open core model making sense. It’ll gradually make more sense.
The counter-argument to that is like, “There’s only going to be a finite number of people writing code. The AI and the ML, Machine Learning, systems are going to be writing code for us. This is going to get automated away.” I tend not to believe that but there are certainly lots of counter-arguments to what I’m saying. The consequence if my thought experiment turns out to be correct is the maximization of this empathy that software creators have for product design. Right now, software creators are very deeply technical in nature. They build libraries, frameworks, plumbing, and infrastructure.
Over time, as we have a billion humans, which is a significant chunk of our species on Earth, creating software, what will happen from there is hugely increased empathy. When those people have empathy for all these open fundamentals, we’re going to see a lot of COSS, open core businesses, and platforms that are very different than the ones we see now. Just because of the increased empathy. I could be wrong about all that. I could be totally insane, but that’s my theory.
I think the intent of the COSS company comes out in the license that they ultimately choose. There are different levels of this empathetic company building or positive-sum examples, and the license says a lot about what the company is setting out to do. Do you think in that same time that we are going to evolve the way we think about licensing as well? I realize licensing is changing every day, but it seems like to reach that in state that you’re speaking about, you’d have to have a lot of the COSS companies being very altruistic in their choice of how they license to achieve that outcome.
It’s a bit different than altruism but I would agree with you on the thought that we’re going to see lots of licensed changes and evolutions. That’s definitely true. There’s going to be a lot of innovation in open source licensing as well as in what’s referred to as source available licensing. Things like the SSPL or licenses that are almost all open source except for a couple of critical distinctions that ensure cloud providers can’t directly compete, commercialize the software, or other competitors. We’re going to see a lot of evolution there. One of my partners, Heather Meeker is the author and an originator of most of these licenses. Ben had her on his show.
The license changes definitely do creep into these thought processes and they do serve to change some of the ground rules, but at the end of the day, if you look at binary package software, fully proprietary, delivered as a service like the whole SaaS industry, a $2 trillion category. There’s a huge difference between doing that and progressively going and saying, “All of our software source is available.” Unity does this, for example. They have a much more restrictive license. You can’t read the code. You can’t even contribute to it. It’s a big, hairy C# code base and a very proprietary company, but there’s a progression.
In anything sufficiently complex, there’s a big gradient and a big spectrum of tradeoffs. Open source is somewhat similar to that. There’s like copyleft, medium copyleft, and neutral copyleft. Copyright is out of the equation but there’s a spectrum of permissiveness in open source licenses. There are about 100 open source licenses. We should see some standardization there. In the source available space, there are increasingly like 100 plus source available licenses which is also this Cambrian explosion. I don’t think that it’s going to dramatically hurt or change a lot of the fundamentals of positive-sum aspects that I was hinting at. There’re certainly different tradeoffs in terms of, for example, the difference between packaged software that’s totally closed binary, and something that’s source available. It’s really profound.
On one hand, you can’t see anything. You have to pay to receive any value. On the other hand, you can see the code and you can contribute if you want. Maybe there’s an option there, maybe not. You can understand how something’s implemented. You need to write or implement like a reverse engineering tool to understand what’s going on. It’s a massive improvement for companies even to embrace source available versus open source.
A bunch of companies are still worried, scared, frightened, and concerned about, “How are they going to differentiate? How are they going to actually capture value and make money if they open-source their core?” I’ve probably talked with a couple of dozen closed core companies over the years that have asked like, “JJ, how do we do this? We’re thinking of going open source because we’re slowing down our growth and we’re not meeting our milestones. We want to open source and we think that’s the magic bullet that’s going to get us to the promised land.” My usual answer to that is like, “Not really. That’s probably not smart because you’re going to have to set expectations with your stakeholders all over again about your new modality.” You can’t expect this to solve your problems. Open source tends to be effective when it’s extremely authentic and genuine in nature in terms of how it originates. When it’s fabricated and contrived, it tends to be less interesting and not as successful.
One of the things that I’ve found interesting talking to all these founders is when I asked them why they chose their license. No one can answer that question. It’s almost like spiritual guidance. They’re like, “This felt like the one.” “I read them all and this was the obvious one,” but they can’t enunciate why. It’s like a gut feel, which I found interesting. I went through the same process. I was like, “Of course, I don’t want to be able to do this, but I want to be able to do that.” I don’t know why I thought that. It was just in my head.
The motivations for license choices are definitely a little opaque sometimes.
Do you think there’s more innovation that’s waiting to be found around that licensing?
I think so. My points of view on open source are pretty opinionated. Open source means something very specific. It’s quite a sacred thing and it builds on Richard Stallman’s four freedoms that were reframed and rebranded, but maintained and recodified with the creation of open source in 1998. Since then, we’ve had some shift and change away from extreme copyleft or strong copyleft licenses, whether it’s GPLv3, the Affero GPL, or similar licenses to things more like Apache 2.0, MIT, or BSD for different reasons. I don’t think open-source, in terms of the core tenants building on the four freedoms, building on what open source popularized in, and democratized for business in ’98, which is the purpose of creating the open-source brand is going to change even over the next decades. I don’t believe the cloud merits redefining open source at all. Many people believe that in the venture community, I think they’re wrong.
Many friends of mine very strongly believe that. I don’t say, “You’re no longer my friend because you don’t agree with me on this.” I just think they’re wrong. There’re two problems in terms of the choices people have, source available licenses, and open source licenses. On both sides, there are too many choices which causes unclear thinking in terms of how people make decisions. On GitHub, it’s pretty easy to go and reach for like, “It’s suggesting Apache 2.0, I’ll pick that, no problem. I’ll slap it on the repo.” That’s what most people do. There’s nothing wrong with that but there’s also a lot of people that dig into that question like, “I want to be more intentional about the license choice.”
When they go and walk down that path a little further, they’re confronted with 100 choices like 100 licenses which is the number of open-source licenses out there that are approved by the OSI. On the other side, for source available, there’s a similar number increasingly of source available licenses with all kinds of different tradeoffs, motivations, interests, and so on. What I’ve talked a lot about with Heather and Bruce, as well, it’s another one of my partners who is a cofounder of open source with Eric Raymond in 1998 and a handful of other people. We need standardization on both sides of the political aisles, if you will, like source available and open source. We need fewer choices and we need more standardization. Maybe ways of having extensible tooling.
Heather, in her great wisdom and vision along with a handful of other folks in the legal community formed a project called PolyForm, which are tools and building blocks for discovering and understanding which set of constraints for source available licensing. It’s not open source but source available licensing could make sense for you. A lot of people are starting to adopt that. Heather also drafted and helped write the SSPL license which MongoDB and Elastic have standardized along with other options and licenses as well that are source available. The Elastic community license and the MongoDB uses SSPL. We need more standardization. I don’t think there’s any perfect answer to what kind of standardization but fewer choices would be useful.
Are there any licenses that you guys have not invested into from an approach perspective and a team? Would anything qualify someone out for you guys as a fund?
Not really. In terms of our beliefs around open source, we’ve passed on investing in companies that self-identify as “we’re a COSS company” but from the get-go, from the very beginning, they’re using a source available license for everything and that’s all they use. They’re not using an open-source license, which means something very specific. We’ve passed on that basis because licenses, from the very beginning, I think there’s a critical need to use a much more permissive open source license. You can use something that’s more copyleft in nature that has less permissiveness like the Affero GPL license, for example. It’s very similar to the SSPL with some minor changes. That’s considered as an open-source license but starting with something that’s more restrictive like the BSL, the Business Source License, which is another license that Heather Meeker adopted. Maybe that’s splitting hairs because it’s proprietary but then it’s open source after a certain timeframe for a year after the latest release or something like that.
We’re still evolving our thinking but the main thesis constraint for us is we invest in COSS companies and COSS companies are building around in open-source core, which means something specific. It’s easy to suss out the motivations and intentions of founders like, “Are they building something to have a stakeholder ecosystem that is vibrant, diverse, and inclusive?” By the way, it is orthogonal to whether or not you want to optimize for external contributions. Each one of these things is somewhat orthogonal. There are lots of things around open source that you can choose to embrace and other things that you can choose to not embrace. That doesn’t mean that you’re less open source than someone else. The community dynamics are somewhat orthogonal to the license choice. In summary, we have a clear enough and well-defined thesis that it’s easy for us to go and understand how to make decisions.
What’s running a conference during the global pandemic like?
That was hard and tricky. It is a lot easier than running a physical conference. I’ll tell you that. The first open core summit was in September 2019 in San Francisco, just before the pandemic. We had around 1,000 people register and a little under that physically attend. It cost $1 million and we had to spend $200,000 on food. A year later, we had 10x more attendees, but our biggest cost item was paying for Hopin licensing. We didn’t have to feed people. Still a lot of work, coordination, and complexity. Mostly communicating with the 250 plus speakers, tens of sponsors, and then coordinating all the stuff, so a lot of emails.
Upwards of 5,000 emails just on the conference with a very small team, it’s definitely a lot of work but doing digital events is way easier. We’re going to be announcing that the Open Core Summit 2021, which will happen in November 2021 is also going to be digital. I don’t think that hybrid or physical events of scale are going to happen in 2021. I’m a lot more optimistic about that for 2022, but not for 2021.
I’m old enough to be recycled out of the original dot-com boom back in the late ’90s and early ’20s. The thing that I remember from that is there was a bunch of people in the industry who didn’t care about it at all. They were just running after the money. That was part of why there were a lot of problems with the original dot-com, but that definitely didn’t help. Do you worry that a similar thing might happen here where something that is quite from deep within you in terms of your thoughts of open source and making things available to other people? Is there a danger that there could be a gold rush and it tramples over the entire segment?
I don’t think that will happen. I don’t think there’s going to be any trampling over the entire segment due to rationality in chunks or parts of the founder ecosystem building these types of companies. I do, however, believe that to use a John Grisham intense analogy, a reckoning is coming. I blogged about this. I thought it was a funny blog. By my count, we have about 100 COSS companies now that have more than $5 million or so in funding, or maybe several million in funding. A great deal of those companies did not exist a few years ago, almost all, frankly. Most of them have no more than several hundred stars on GitHub and a pitch deck. A lot of those companies are going to face hard questions around how they continue to build towards the outcomes that their investors are expecting, which is billion-dollar outcomes.
There’s going to be a lot of incredible valuable companies to come out of that but I also think that founders have to be super careful not to hyper optimize for the extreme overabundance of money in the world in all asset classes. In particular, venture capital on private startup financings and focus on making their company work. At the end of the day, if you raise money from investors around open-source technology, you have to build a business, and you have to figure out how to generate revenue, build a product, go to market, and hire amazing people.
If you fail, you fail. It’s not the end of the world and you learned a lot, but I do think that there’s a lot of funding for companies that shouldn’t have raised funding or companies that got overfunded when they raised. They should have raised $1 million or $2 million instead, they raised a $10 million round or a $15 million round. I definitely think there’s a lot of that. It’s a function of far smarter people than me. I don’t consider myself that smart and I said, “There’s too much money chasing too few opportunities in the world and that creates a supply-demand effect that is suboptimal for a lot of stakeholders.”
Have you got anyone you want to shout out to or thank for anything in the last year?
I would say shout out to the open-source community for continued commitment and passion to build open technology. There are small fractions of that ecosystem with loud voices counteracting a lot of the positive-sum aspects. They’re the loud minority. On the whole, there’s an incredibly genuine and positive-sum-driven way of building out there, and that’s representative of the vast majority of the open-source world. Without those people and those voices, none of what I’m working on would be possible. It’s also true. None of most of the tech industry would be possible, either. Not one particular individual, but that collection of people.
I’m hoping that in the near future, I’ll be able to get on a plane and fly to San Francisco and join the 2021 but it might have to be 2022 or you could host it in London if you wanted.
We’ll eventually do country-specific events but we have to get past this pandemic. For now, we have a lot of great internet-based technologies that bring a lot of people together. I look forward to meeting with you in person at some point in the future.
I want to make that happen. Thanks for your time.
Joseph Jacks is the Founder and General Partner of OSS Capital. He is based in San Francisco.
Previously, Joseph was an EIR at Quantum Corporation in support of the Rook project which was subsequently donated to the CNCF (its first storage project) where he has also been involved (at the board level, in various committees and as an advisor) since inception. Joseph also founded KubeCon (now also run by the Linux Foundation’s CNCF) while also founding and building Kismatic, the first enterprise-focused commercial Kubernetes company (acquired by Apprenda, subsequently acquired by ATOS). Over the preceding several years, Joseph worked at Mesosphere (now D2IQ), Enstratius (acquired by Dell Software), TIBCO Software and Talend in various sales, engineering, product and strategy capacities.
Joseph is also the founder of Open Core Summit, the global COSS ecosystem conference, run independently as a vendor-neutral community that includes the leading investment firms, cloud providers, COSS companies, analysts, F2000 enterprises and more. The first OCS occurred in San Francisco, September 19-20th, 2019, uniting nearly 1,000 attendees across 29 countries, 40 major sponsors and 100~ speakers. The second annual OCS 2020 occurred fully online from December 16-18th, 2020, with 13,000+ attendees, 200~ speakers and 50~ major sponsors, celebrating and democratizing knowledge about the meteoric growth of the COSS category.